Recent Advances in Volatile Sulfur Compounds in Cheese: Thiols and

with the flavor of different foods, including fermented ones. ... It was interesting to note that the data also showed that other. S-methyl .... the c...
3 downloads 0 Views 1MB Size
Chapter 6

Downloaded by EAST CAROLINA UNIV on November 5, 2016 | http://pubs.acs.org Publication Date (Web): August 24, 2011 | doi: 10.1021/bk-2011-1068.ch006

Recent Advances in Volatile Sulfur Compounds in Cheese: Thiols and Thioesters A. M. Sourabié,a,b H.-E. Spinnler,a A. Saint-Eve,a P. Bonnarme,a and S. Landaud*,a aAgroParisTech/INRA,

UMR 782 GMPA, Rue Lucien Brétignières, 78850 Thiverval-Grignon, France bSAF-ISIS, Zone Artisanale, 40140 Soustons, France *E-mail: [email protected].

Among the numerous compounds involved in cheese aroma, sulfur compounds are of particular interest because of their very powerful odors and low perception thresholds. Nevertheless, little attention has been focused until now on the possible presence of polyfunctional thiols in cheese, even if these compounds have been found to be associated with the flavor of different foods, including fermented ones. The difficulty in isolating volatile thiols in cheeses is probably due to the complexity of this matrix and to the very low concentrations of these compounds that make them undetectable using classic means. Consequently, different methods of sample preparation followed by extractions with p-hydroxymercuribenzoate were used to investigate the possible occurrence of thiol compounds in ripened cheeses. The analysis of cheese extracts by GC coupled with pulse flame photometry, MS and olfactometry detections made it possible to identify ethyl 2-mercaptopropionate (ET2MP) and ethyl 3-mercaptopropionate (ET3MP) in smear and mold-ripened cheeses for the first time. The presence of ET3MP in cheeses at concentrations of around 3 µg/kg, significantly higher than its perception threshold in cream (723 ppt), suggests that this thiol may significantly contribute to the aroma of these cheeses. Concerning other potent odors that contribute to the aroma of numerous cheeses, we also investigated the ability of Brevibacteria strains to produce S-methyl thioesters in © 2011 American Chemical Society Qian et al.; Volatile Sulfur Compounds in Food ACS Symposium Series; American Chemical Society: Washington, DC, 2011.

Downloaded by EAST CAROLINA UNIV on November 5, 2016 | http://pubs.acs.org Publication Date (Web): August 24, 2011 | doi: 10.1021/bk-2011-1068.ch006

the presence of (i) methanethiol, (ii) fatty acids, or (iii) branched-chain amino acids as precursors. All the strains studied were able to yield the corresponding S-methyl thioesters (i.e., S-methyl thioisovalerate from L-leucine) from precursors. It was interesting to note that the data also showed that other S-methyl thioesters e.g., S-methyl thioacetate or S-methyl thioisobutyrate, were also produced following the addition of an individual precursor (e.g., L-leucine). Enzymatic and tracing experiments allowed us to propose the catabolic pathways used by the strains to produce S-methyl thioesters. Keywords: smear soft cheese; polyfunctional thiols; S-methyl thioesters; Brevibacteria; fatty acids; branched chain amino acids

Introduction Cheese ripening is a complex phenomenon in which a wide variety of microorganisms (e.g., yeasts, bacteria) are involved. Their action results in the synthesis of a variety of aroma compounds, especially volatile sulfur compounds (VSC), which give specificity to a variety of ripened cheeses. Even if the VSC found in cheeses belong to a large number of chemical families (1), we chose to focus on thiols and thioesters because little is still known about their nature (thiols) and their biosynthesis (thioesters). Little attention has been given to the possible occurrence of thiols in cheese until now. Only methanethiol (MTL) and H2S, which are common precursors for a variety of other VSC, have been frequently reported in cheeses (Table I). MTL, found in high amounts in Camembert cheese, gives a characteristic “cooked cabbage” flavor note. It has also been reported in vintage Cheddar, Parmesan, Pecorino, Grana Padano and blue cheese. Among those cheeses, MTL was perceived as a “strong aroma” in Cheddar and blue cheese, and as a “very strong” aroma in Parmesan (3). MTL is the first product of methionine degradation. Moreover, H2S has the unpleasant odor of “rotten eggs” and has been reported in Limburger cheese (Table I) but only in traces, probably because of its high reactivity and its difficulty to be quantified. H2S is considered as the primary degradation product of cysteine. Few other thiols have been reported in cheeses to date. 4-mercapto-4-methylpentan-2-one has been reported in Gouda cheese as a catty flavor compound (4). More recently, Kleinhenz carried out studies that highlighted the hypothetical presence of several thiols in Cheddar cheese (5). However, they used a phosphine reagent, TCEP (tris(2-carboxyethyl) phosphine) to prevent the oxidation of thiols and to enable their recovery from Cheddar cheese oil. Given that TCEP is a very powerful reducer, the hypothetical thiols that they found could have been formed by the reduction of the polysulfur molecules. Moreover, these authors were not able to unequivocally characterize the identified thiols. 120 Qian et al.; Volatile Sulfur Compounds in Food ACS Symposium Series; American Chemical Society: Washington, DC, 2011.

Downloaded by EAST CAROLINA UNIV on November 5, 2016 | http://pubs.acs.org Publication Date (Web): August 24, 2011 | doi: 10.1021/bk-2011-1068.ch006

Table I. Examples of thiols found in cheese (2)

a

Thiol compounds

Flavor note

Odor threshold (ppb)

Probable precursor

Cheeses in which they occur

Hydrogen sulfide

Rotten eggs

0.18a

Cysteine

Limburger Cheddar

Methanethiol

Cooked cabbage; fermented cabbage

0.06b

Methionine

Camembert Cheddar blue cheese Parmesan Grana Padano Pecorino

4-mercapto-4methyl-pentan2-one

catty

nd

Gouda

In air.

b

In sunflower oil.

c

nd

c

nd: not determined.

Nevertheless, thiols possess a wide variety of odors ranging from cheese to blackcurrant, depending on their chemical structures (6) and their concentrations, and have been found to be associated with the flavor of different foods, including fermented ones (7–10). Because of the multiple sulfur descriptors usually used to characterize smear soft cheeses like Munster, it is reasonable to assume that thiols may be involved in their characteristic flavors. The difficulty in isolating volatile thiols in cheese is due to the complexity of the cheese matrix and to the very low concentrations of these compounds, which make them undetectable using classic means. In the first part of this chapter, we present the adaptation of a thiol extraction method to the cheese matrix, allowing us to detect and identify two new thiols in cheeses. The hypothesis concerning their biosynthesis is also developed. Once formed, certain thiols such as MTL can be oxidized to form other VSC such as sulfides and thioesters. S-methyl thioesters have been reported in several cheeses and extensively studied with respect to their detection thresholds—which ranged from 1 to 3 ppb—and flavor notes (Table II). The most common descriptors cited for thioesters were “cabbage”, “garlic” and “cheesy” (11). For example, S-methyl thioacetate (MTA) was detected in Vacherin, Pont-l’Evêque, Langres and Epoisses; S-methyl thiopropionate (MTP) was found in Vacherin and Pont-l’Evêque (Table II). The production of S-methyl thioesters, particularly those with a carbon chain length of 2-6 carbons, has been extensively studied over the years. Hence, it was shown that strains of Geotrichum candidum were able to produce S-methyl thioesters in a liquid cheese medium (12, 13) and that Micrococcus (14) and Brevibacterium strains were able to produce S-methyl thioacetate (15). The production of these volatiles was achieved with cell-free extracts of G. candidum and their synthesis was shown to be essentially spontaneous (16). Furthermore,

121 Qian et al.; Volatile Sulfur Compounds in Food ACS Symposium Series; American Chemical Society: Washington, DC, 2011.

Downloaded by EAST CAROLINA UNIV on November 5, 2016 | http://pubs.acs.org Publication Date (Web): August 24, 2011 | doi: 10.1021/bk-2011-1068.ch006

Lamberet et al. (17) succeeded in generating various S-methyl thioesters by resting cells of coryneform bacteria including Brevibacterium strains. S-methyl thioesters are believed to originate from MTL and acyl-CoAs (14, 18) that could be generated by numerous metabolic pathways such as those of sugar, fatty acids and branched-chain amino acids. During cheese ripening, Brevibacterium linens, a well-known cheese-ripening bacterium, is able to produce methanethiol from methionine degradation (19) during cheese ripening via a methionine γ-lyase (20, 21), and acyl-CoAs probably intermediates in catabolic pathways involving FAs and BCAAs. After examination of the genome sequence of B. linens, it was in fact shown that this species may have the ability to synthesize FAs and branched-chain FAs from BCAAs (22) and could therefore generate acyl-CoAs in a series of reactions initiated by aminotransferases (ATases). These enzymes, which were previously reported in B. linens (23, 24), catalyze the first step of amino acids catabolism, converting them into cheese flavor compounds. In addition, based on the same genomic data, it has been reported that this species possesses all the genes required for the degradation and biosynthesis of FAs (22). However, despite these findings, the mechanisms and biosynthetic routes for the production of S-methyl thioesters in cheese have not yet been conclusively established.

Table II. Examples of thioesters found in cheese (2)

a

Thioesters

Flavor note

Odor threshold (ppb)a

Probable precursor

Some cheeses in which they occur

S-methyl thioacetate

Cabbage, cheesy, crab

3

MTL and acetyl-CoA

Pont-l’Evêque Langres Epoisses Vacherin Limburger

S-methyl thiopropionate

Cabbage, cheesy, garlic, crab

2

MTL and propionylCoA

Vacherin Pont-l’Evêque

S-methyl thiobutyrate

Cabbage, cheesy, rancid, garlic

3

MTL and butyrylCoA

S-methyl thioisovalerate

Cheesy, garlic, cabbage

1.2

MTL and isovalerylCoA

S-methyl thioisobutyrate

Garlic, cheesy, cabbage

2.6

MTL and isobutyrylCoA

In water.

122 Qian et al.; Volatile Sulfur Compounds in Food ACS Symposium Series; American Chemical Society: Washington, DC, 2011.

Consequently, in the second part of this chapter, we demonstrate the ability of Brevibacterium strains to produce S-methyl thioesters from FAs and BCAAs and to elucidate the metabolic pathways associated with their synthesis.

Polyfunctional Thiols Are Involved in Cheese Aroma

Downloaded by EAST CAROLINA UNIV on November 5, 2016 | http://pubs.acs.org Publication Date (Web): August 24, 2011 | doi: 10.1021/bk-2011-1068.ch006

Identification of Two New Thiols in Cheese A triangular sensory test was firstly performed to evaluate the possible impact of thiols on overall aroma of cheese. To do this, smear-ripened cheese (Munster), alone or mixed with p-hydroxymercuribenzoate (pHMB), was tested using the well-known property of thiols to react with mercury. This chemical bond normally leads to the loss of the sensory properties of thiol. Due to the fact that the sensory difference between the two samples, with or without pHMB, was clearly significant and that the flavor descriptors used for samples with pHMB were solvent, ammonia and white cheese, whereas those of the controls were fruity, sulfured and cheese, we assumed that thiols were key compounds responsible for aroma in smear-ripened cheese. These results may have been due to thiols such as hydrogen sulfide or methanethiol, which have been reported to contribute to cheese flavor. However, given that these two thiols are very reactive and that cheese flavor is a complex mixture of a wide range of compounds, other thiols was purchased. Consequently, we attempted to isolate these compounds in cheese using a method adapted from that of Tominaga (25, 26). The preparation of samples was considerably modified to overcome the problems associated with the fat content and heterogeneity of cheese. Consequently, three methods were evaluated (27) and parameters such as length and the number of the steps were taken into consideration in the overall extraction efficiency. Only the protocol that considers the thiol content of the cheese surface layers made it possible to eliminate the fat content drawback. This procedure gave the best results for isolating the compound and was therefore selected (see Figure 1 for experimental details). Given that flavor formation in cheese often begins at the surface where microbial activity is the highest before diffusion into the core, this result is not surprising. Moreover the layers were wide enough to include a small part of the cheese core, which probably improved extraction efficiency. Thiols could be identified in 40% of the cheeses analyzed, including the smear and mold-ripened ones. Using two specific and sensitive detectors (a Pulsed Flame Photometric Detector (PFPD) and Mass Spectrometry (MS) coupled with Gas Chromatography (GC)), thiol identification was based on their retention times, PFPD indices (27) and mass spectra in comparison with commercial references. One of them did not match any compound in the GC-MS library (NIST Mass spectral database), perhaps because this detector is not sensitive enough and undoubtedly because the concentration of the compound was too low for detection. Nevertheless, it has been tentatively identified as ethyl 2-mercaptopropionate (ET2MP) by comparison to its PFPD retention time, PFPD index and those of the commercial reference. This volatile is known to be found in apple juice and strawberries (28, 29). 123 Qian et al.; Volatile Sulfur Compounds in Food ACS Symposium Series; American Chemical Society: Washington, DC, 2011.

Downloaded by EAST CAROLINA UNIV on November 5, 2016 | http://pubs.acs.org Publication Date (Web): August 24, 2011 | doi: 10.1021/bk-2011-1068.ch006

Figure 1. Protocol for thiol extraction from the cheese matrix. The cheese surface was cut into pieces before dichloromethane extraction.

Figure 2. Mass spectra (MS/EI) of Ethyl 3-mercaptopropionate detected in a final L-cysteamine extract of Munster cheese (A) compared to that of its commercial analogue (B). A second sulfur compound was identified as ethyl 3-mercaptopropionate. The mass spectrum of this compound (Figure 2A) closely matched that of the commercial reference (Figure 2B). The calculated linear retention index of ET3MP was also identical to that of the reference (on a DBXLB column). In addition, ET3MP identification was confirmed in SIM mode by overlapping selected ions, m/z 61, 88, and 134, at the linear retention index of the reference compound. Moreover, an additional attribute for ET3MP identification was represented by the 124 Qian et al.; Volatile Sulfur Compounds in Food ACS Symposium Series; American Chemical Society: Washington, DC, 2011.

odor quality of this compound (assessed by Gas Chromatography-Olfactometry), which was identical to that of the commercial reference at the same concentration. Ethyl 3-mercaptopropionate was previously reported in various foods such as wine (30) or Concord grape (31). To our knowledge, however, this is the first time that it has been found in cheese.

Downloaded by EAST CAROLINA UNIV on November 5, 2016 | http://pubs.acs.org Publication Date (Web): August 24, 2011 | doi: 10.1021/bk-2011-1068.ch006

Evidence for an Unequivocal Impact of ET3MP in Cheese Aroma To evaluate the impact of ethyl 3-mercaptopropionate on the overall aroma of cheese, the concentration of this compound was quantified using the standard addition procedure. The concentrations for ET3MP were approximately 2 μg/kg and 4 μg/kg for a mold-ripened cheese and a smear-ripened cheese, respectively (standard deviation ± 0.2 μg/kg). Furthermore, the perception threshold of this compound was measured in fresh cream using triangle tests with ascending flavor concentrations. The group perception threshold for detection of ET3MP in fresh cream, as detected by panelists (geometric mean of the individual thresholds), was 723 ppt, evidence of its very powerful odor (Figure 3). The presence of ET3MP in cheese at concentrations higher than its perception threshold in cream suggests that this thiol may play a significant role in the aroma of these cheeses.

Figure 3. Number of panelists with a correct answer to triangle tests with ascending concentrations of ET3MP in fresh cream. Key: a, group BET was the geometric mean of the individual BET; b, critical number (minimum) of correct answers required for significance at the stated significance level (5%). 125 Qian et al.; Volatile Sulfur Compounds in Food ACS Symposium Series; American Chemical Society: Washington, DC, 2011.

Downloaded by EAST CAROLINA UNIV on November 5, 2016 | http://pubs.acs.org Publication Date (Web): August 24, 2011 | doi: 10.1021/bk-2011-1068.ch006

Hypothetical Metabolic Pathways of These New Thiols in Cheese The metabolic routes leading to the formation of both ET2MP and ET3MP in cheese have not yet been elucidated and may be relatively complex. However, it can be assumed that they are quite similar to those described for other thiol compounds in wine and beer. ET2MP and ET3MP synthesis could therefore be related to the microbial catabolism of amino acids, mainly alanine, cysteine and methionine, which are present in caseins. Regardless of the pathway, it is likely that 2- and 3-mercaptopropionate were formed before subsequent esterification with ethanol. Alanine could first be converted by an ammonia lyase to synthesize prop-2-enoic acid (acrylate). Acrylate could then react with hydrogen sulfur (H2S) derived from cysteine to yield either ET2MP or ET3MP in a Markovnikov-type reaction (Figure 4). Moreover, alanine could undergo a dehydrogenation catalyzed by an alanine dehydrogenase (EC 1.4.1.1) to produce pyruvate, which would finally yield 2-mercaptopropionate by a nucleophilic substitution with H2S (Figure 4). Alternatively, pyruvate could be formed by glycolysis during cheese ripening. Furthermore, since methionine can produce homocysteine by a methyl transfer, an additional possible route for the formation of ET3MP could be the Ehrlich degradation of this amino acid, which is known to occur in fermented beverages and cheese.

Figure 4. Hypothetical formation pathways of ethyl 2- and ethyl 3-mercaptopropionate. 126 Qian et al.; Volatile Sulfur Compounds in Food ACS Symposium Series; American Chemical Society: Washington, DC, 2011.

Downloaded by EAST CAROLINA UNIV on November 5, 2016 | http://pubs.acs.org Publication Date (Web): August 24, 2011 | doi: 10.1021/bk-2011-1068.ch006

Conclusion: Could Hydrophobic Thiols Be Efficiently Extracted from the Cheese Matrix? To conclude, the promising results presented here are a first step towards a better understanding of thiol contribution to cheese flavor. Nevertheless, the gas-liquid partition coefficient (measured using the Phase Ratio Variation (32)) of ET3MP was 1.75 10-3 in water and four times less in cream (4.87 10-4), showing that this volatile is only moderately retained in fresh cream. This result is not surprising since the molecule is slightly hydrophobic (log P = 1.40). Consequently, identification of more hydrophobic thiols in cheese will require improvements of the extraction method. For example, the use of more apolar solvents (first step of extraction) or the use of “AFFi-Gel” (PhHgloaded Agarose gel) for purification and enrichment of volatile thiols could be tested. Finally, the fact that ET3MP was only found in certain types of cheeses, and not in all of the tested samples strongly suggests that the microbial flora in the ripening chamber must be involved in its synthesis. Hence, differences between cheese-making environments must be taken into account when assessing cheese samples from the same variety.

Involvement of Branched Chain Amino Acid (BCAA) and Fatty Acid (FA) Catabolism in the Biosynthesis of S-methyl Thioesters Requirement of an Enzymatic Step in the Synthesis of S-methyl Thioesters To determine the involvement of microorganisms in the pathways leading to S-methyl thioesters, we performed preliminary experiments using different precursors without microbial cells: acyl-CoAs, methanethiol (MTL), dimethyldisulfide (DMDS), FAs or BCAAs. The results obtained showed that in the absence of cells, thioesters were formed only in the mixtures containing acyl-CoAs and MTL, and that their synthesis was not significantly enhanced in the presence of resting cells. Consequently, S-methyl thioesters are formed by a spontaneous reaction between acyl-CoAs and MTL. Considering the chemistry of thiols, it is assumed that a very reactive compound such as MTL replaces the large thiol molecule, coenzyme A, via a nucleophilic substitution reaction. In contrast, since FAs and BCAAs only generate thioesters in the presence of cells, we suspect that their activation to thioester precursors might occur via an enzymatic reaction to yield acyl-CoAs. Consequently, we studied two species of Brevibacteriaceae (B. antiquum and B. aurantiacum) as major actors in cheese ripening, particularly via their ability to produce volatile sulfur compounds (VSCs). Although significant differences in thioester concentrations were observed, both species exhibited very similar thioester production patterns.

127 Qian et al.; Volatile Sulfur Compounds in Food ACS Symposium Series; American Chemical Society: Washington, DC, 2011.

Downloaded by EAST CAROLINA UNIV on November 5, 2016 | http://pubs.acs.org Publication Date (Web): August 24, 2011 | doi: 10.1021/bk-2011-1068.ch006

Involvement of BCAA and FA Catabolism in S-Methyl Thioester Synthesis As shown in Figure 5, the addition of a specific BCAA (leucine or valine) or a specific FA (propionate, butyrate or isovalerate) led to a significant increase in the corresponding thioester: (i) S-methyl thioisovalerate (MTiV) for leucine or isovalerate; (ii) S-methyl thiopropionate (MTP) for propionate; (iii) S-methyl thiobutyrate (MTB) for butyrate; and (iv) S-methyl thiosiobutyrate (MTiB) for valine. Utilization of labeled precursor compounds (the precursor was totally substituted with the labeled precursor for tracing experiments) coupled with GS-MS for thioester analyses enabled us to confirm the strong connection between the precursor added and the formation of the corresponding thioesters. Concerning the addition of a specific FA, labeled propionate, butyrate or isovalerate generated the corresponding labeled S-methyl thioester. The labeling efficiency (calculated according to the method of Arfi et al. (33)) was always higher than 85%, regardless of the thioester. The small proportion of unlabeled compounds may have been derived from intracellular pools of FAs, amino acids or sugars, as suggested by the controls during thioester biosynthesis assays. Depending on the structure of the labeled precursor (Table III), different types of labeling were observed by GC-MS.

Figure 5. Production of S-methyl thioesters (in mMole) from propionate (C3), butyrate (C4), isovalerate (iC5), L-leucine (Leu), L-valine (Val) and without the addition of precursors (control: cells in plain buffer) by whole cells of B. antiquum incubated for 5 h with methanethiol. MTA (S-methyl thioacetate); MTP (S-methyl thiopropionate); MTB (S-methyl thiobutyrate); MTiB (S-methyl thioisobutyrate); MTiV (S-methyl thioisovalerate). Distinct letters (e.g., a, b, c and d) were assigned to significantly different groups (one-way analysis of variance and Newman-Keuls (P < 0.05) tests). For example, the mass spectra of labeled MTiV produced from [1-13C] isovalerate showed a molecular ion at m/z=133 (Figure 6-B) that corresponded to a mass increase of M+1 compared to the unlabeled molecule (molecular ion 128 Qian et al.; Volatile Sulfur Compounds in Food ACS Symposium Series; American Chemical Society: Washington, DC, 2011.

at m/z= 132) (Figure 6-A). Since the labeled precursor had one 13C in the first position, the labeled fragment in MTiV may have been derived from this part of the [1-13C] isovalerate. This M+1 increase was also observed for labeled MTB while a 2-mass unit increase was observed for labeled MTP, corresponding to the two deuterium atoms of its labeled precursor, [D2] propionate.

Downloaded by EAST CAROLINA UNIV on November 5, 2016 | http://pubs.acs.org Publication Date (Web): August 24, 2011 | doi: 10.1021/bk-2011-1068.ch006

Table III. Structures of labeled compounds used in the study of thioester biosynthesis Labeled molecule

Structure

[D2] propionate

CH3CD2COOH

[1-13C] butyrate

CH3CH2CH213COOH

[1-13C] isovalerate

(CH3)2CHCH213COOH

[D10] L-leucine

(CD3)2CDCD2CDNH2COOH

[D8] L-valine

(CD3)2CDCDNH2COOH

Figure 6. GC-MS spectra of unlabeled S-methyl thioisovalerate (MTiV) (A) and labeled MTiV derived from [1-13C] isovalerate (B) and [D10] L-leucine (C). The labeled MTiV molecules exhibit a molecular mass increase of m+1 and m+9, respectively, compared to the unlabeled molecule (A; molecular mass at m/z=132). 129 Qian et al.; Volatile Sulfur Compounds in Food ACS Symposium Series; American Chemical Society: Washington, DC, 2011.

Downloaded by EAST CAROLINA UNIV on November 5, 2016 | http://pubs.acs.org Publication Date (Web): August 24, 2011 | doi: 10.1021/bk-2011-1068.ch006

In light of these results, it is possible to suppose that when added to the mixtures, label-free FAs are activated by an acyl-CoA synthase to form acyl-CoAs before reacting with MTL to synthesize thioesters. The mechanism of this reaction using ATP and free CoA could be schematized as follows:

The trans-acylation reaction is known to be the first step in FA catabolism. This reaction is catalyzed by acyl-CoA synthases. These enzymes are members of the so-called acyl-adenylate/thioester-forming family. Concerning thioesters derived from BCAAs, spiking of labeled L-leucine and L-valine led to the generation of the corresponding labeled thioesters (MTiV and MTiB for L-leucine and L-valine, respectively), with a labeling efficiency close to 90%. The mass spectra of labeled MTiV (Figure 6-C) and MTiB (data not shown) showed molecular ion mass increases of M+9 and M+7, respectively, compared to the unlabeled molecules (molecular ions at m/z=132 and m/z=118, respectively). The analysis of the mass spectra of labeled compounds indicated that they could be derived from labeled isovaleryl-CoA and isobutyryl-CoA, respectively. The type of labeling observed with MTiV perfectly matched the following series of reactions: labeled [D10] L-leucine is assumed to lose one deuterium atom when it is transaminated to α-ketoisocaproic acid (KIC); labeled KIC is subsequently oxidatively decarboxylated to isovaleryl-CoA before yielding MTiV with an M+9 increase, compared to the unlabeled molecule. This observation supports the hypothesis that the studied strains are able to perform transamination of amino acids to α-keto acids. Consequently, we can hypothesize that the conversion of L-leucine and L-valine to S-methyl thioesters involved enzymatic steps and was initiated by aminotransferases (ATases). Once they were formed, the α-keto acids were subsequently converted to acyl-CoAs by a KADHase before reacting with MTL to generate thioesters. Hence, the strains studied exhibit ATase and KADHase activities, as shown by our enzymatic assays (data not shown). The drastic reduction in the synthesis of thioesters caused by the addition of sodium arsenite, a known inhibitor of the α-keto acid/pyruvate dehydrogenase complex, reinforces this view. The branched-chain α-keto acid dehydrogenase is a multienzyme complex that catalyzes the oxidative decarboxylation of branched-chain keto acids with the formation of branched-chain acyl-CoAs and the reduction of NAD+ to NADH. The KADHase complex, which exhibits a broad specificity for BCAAs, was already reported in Bacillus subtilis (34) and Oenococcus oeni (35). It may convert the corresponding keto acids formed by BCAA catabolism to acyl-CoAs. Moreover, this complex was previously shown to be responsible for the production of isovaleryl-CoA from L-leucine in S. cerevisiae (36). Since the concentrations of MTiV and MTiB were significantly enhanced (P < 0.05) in both strains with L-leucine and L-valine utilization, these observations confirm that the strains were capable of producing KIC and KIV before activating them to isovaleryl-CoA and isobutyryl-CoA, which react with MTL to synthesize thioesters. Taken together, these findings and the results of labeling experiments 130 Qian et al.; Volatile Sulfur Compounds in Food ACS Symposium Series; American Chemical Society: Washington, DC, 2011.

Downloaded by EAST CAROLINA UNIV on November 5, 2016 | http://pubs.acs.org Publication Date (Web): August 24, 2011 | doi: 10.1021/bk-2011-1068.ch006

with [D10] L-leucine allow us to propose the hypothetical pathway for L-leucine presented in Figure 7.

Figure 7. Hypothetical pathways of S-methyl thioisovalerate (MTiV) formation from [D10] L-leucine using whole cells of Brevibacterium strains.

131 Qian et al.; Volatile Sulfur Compounds in Food ACS Symposium Series; American Chemical Society: Washington, DC, 2011.

Downloaded by EAST CAROLINA UNIV on November 5, 2016 | http://pubs.acs.org Publication Date (Web): August 24, 2011 | doi: 10.1021/bk-2011-1068.ch006

Is the Great Variety of Thioesters Due to the Combination of FA and BCAA Catabolism and a Concomitant β-Oxydation of acyl-CoAs? As shown in Figure 5, the addition of a specific FA or BCAA significantly increases the production of unexpected thioesters. We therefore investigated the formation of these unexpected compounds (e.g., MTA, MTP or MTB). First, these thioesters were also labeled following supplementation with labeled precursors, showing the link between these volatiles and putative precursors. Three types of labeled MTAs exhibiting molecular mass increases of m+1, m+2 and m+3 compared to the unlabeled molecule were identified in reaction mixtures with [D10] L-leucine. The overall labeling efficiency for the three types of labeled MTA molecules was greater than 45%. Among them, the MTA molecule exhibiting a mass increase of m+3 represents two-thirds of the labeled forms of this thioester. Regarding L-valine catabolism, a [D7] MTiB (labeling efficiency of 90%) and a [D3] MTP (labeling efficiency of 15%) were identified in the samples spiked with [D8] L-valine. This observation provides evidence that L-valine could sequentially be converted into α-ketoisovaleric acid (KIV) and isobutyryl-CoA ,which either generates MTiB or is alternatively degraded to yield propionyl-CoA, the precursor of MTP. FA and BCAA degradations have a common step, which is to generate acylCoAs, leading to the corresponding thioesters (e.g., MTiV with isovalerate or Lleucine). Indeed, in addition to reacting with MTL, acyl-CoAs are also substrates for β-oxidation (Figure 7). Consequently, they could undergo series of reactions that ultimately generate molecules of acetyl-CoA and/or propionyl-CoA. The latter compounds could, in turn, be responsible for the generation of some of the other thioesters (i.e., MTA, MTP and MTB) in reaction mixtures supplemented with a single substrate alone. The general mechanism could proceed as follows. First acyl-CoAs are oxidized to enoyl-CoAs that are, in turn, hydrated to hydroxyacyl-CoAs. The latter compounds are subsequently oxidized in the presence of a free molecule of coenzyme A to ketoacyl-CoAs that undergo a scission catalyzed by a thiolase to release an acetyl-CoA. The remaining acyl-CoAs (shorter by two carbon units) undergo additional cycles of β-oxidation until they are completely transformed to acetyl-CoA, the precursor of MTA. In the case of L-leucine, this molecule is first converted to KIC, which is further activated to yield isovaleryl-CoA, a substrate of the β-oxidation enzymes. These enzymes e.g., acyl-CoA dehydrogenase, enoyl-CoA carboxylase and hydratase, convert isovaleryl-CoA to generate 3-hydroxy-3-methylglutaconyl-CoA (HMG). HMG is subsequently cleaved by the action of an unidentified lyase either into acetyl-CoA or acetoacetate. Acetoacetate is then activated by an acetyl-CoA C-acetyltransferase to acetoacetyl-CoA that, in turn, undergoes the action of an unidentified lyase to generate two molecules of acetyl-CoAs with different types of labeling. These compounds finally yield two labeled MTA molecules with mass increases of m+2 and m+3 compared to the unlabeled molecule. Furthermore, due to the ketoenolic equilibrium, acetoacetyl-CoA derived from L-leucine degradation could undergo 132 Qian et al.; Volatile Sulfur Compounds in Food ACS Symposium Series; American Chemical Society: Washington, DC, 2011.

Downloaded by EAST CAROLINA UNIV on November 5, 2016 | http://pubs.acs.org Publication Date (Web): August 24, 2011 | doi: 10.1021/bk-2011-1068.ch006

dehydration to form butanoyl-CoA that subsequently reacts with MTL to form MTB. The identification of a labeled MTB molecule with a mass of m+4 supports this hypothesis. Examination of the genome sequence of B. linens ATCC 9174 showed that this organism has all the genes required to produce and degrade FAs as previously reported (22). Moreover, the proposed L-leucine degradation pathway could be expanded to L-valine based on the same genomic and genetic data. Thus, L-valine could first be converted to KIV by an ATase and then to isobutyryl-CoA, which subsequently generates MTiB. Isobutyryl-CoA could, in turn, yield 3-hydroxyisobutyryl-CoA via the action of two enzymes of β-oxidation including the previously reported butanoyl-CoA dehydrogenase (EC 1.3.99.2). After the release of coenzyme A, it is assumed that 3-hydroxyisobutyrate is converted into methylmalonate semialdehyde, which finally yields propionyl-CoA and carbon dioxide. All these hypothetical pathways are under investigation in our laboratory using techniques that allow the identification of non-volatile intermediary compounds such as acyl-CoAs.

Towards the Improvement of the Aromatic Quality of Cheeses This study raises the question of the physiological role of S-methyl thioester synthesis. As far as we know, the physiological role of thioester synthesis in Brevibacterium is unknown. Their role can be compared to that of esters in beer and sake yeasts. It has been suggested that ester synthesis could be a useful way for cells to regenerate free CoA in these organisms without releasing high concentrations of free acetic and medium chain FAs that could be toxic (37). Our results provide new insights into the understanding of the formation of these important volatiles that contribute to the aroma of numerous cheeses. We have demonstrated that the catabolism of FAs and BCAAs, the β-oxidation of acyl-CoAs and the catabolism of methionine have to be considered when assessing the pathways leading to the generation of VSCs. The impact of the transport phenomenon for substrates and products as well as the expression and regulation of the genes involved in FA and BCAA catabolism need to be investigated before implementing the production of cheeses with the desired types or balances of S-methyl thioesters.

General Conclusion and Prospects These recent advances concerning key aromatic sulfur compounds will significantly contribute to the better control of their microbial production and to the support of high organoleptic quality of cheeses. Cheese-making results in the combined action of an ecosystem made of yeasts and bacteria, which implies the concept of co-metabolism. The understanding and/or control of the mechanisms of VSC synthesis would therefore remain partial without taking the simultaneity and/or the succession of biochemical reactions generated by the microbial consortium into account. The functional analysis of 133 Qian et al.; Volatile Sulfur Compounds in Food ACS Symposium Series; American Chemical Society: Washington, DC, 2011.

the ecosystem and its interaction with the cheese matrix is therefore essential for the improved control of the full ripening process.

References 1. 2.

Downloaded by EAST CAROLINA UNIV on November 5, 2016 | http://pubs.acs.org Publication Date (Web): August 24, 2011 | doi: 10.1021/bk-2011-1068.ch006

3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23.

Sablé, S.; Cottenceau, G. J. Agric. Food Chem. 1999, 47, 4825–4836. Landaud, S.; Helinck, S.; Bonnarme, P. Appl. Microbiol. Biotechnol. 2008, 77, 1195–1205. Frank, D. C.; Owen, C. M.; Patterson, J. Lebensm.-Wiss. Technol. 2004, 37, 139–154. Badings, H. T.; Stadhouders, J.; Van Duin, H. J. Dairy Sci. 1968, 51, 31–35. Kleinhenz, J. K.; Kuo, C. J.; Harper, W. J. A. Milk Sci. Int. 2005, 60, 275–278. Vermeulen, C.; Pellaud, J.; Gijs, L.; Collin, S. J. Agric. Food Chem. 2003, 49, 5445–5449. Bouchilloux, P.; Darriet, P.; Dubourdieu, D. J. Int. Sci. Vigne Vin 1996, 30, 23–29. Bouchilloux, P.; Darriet, P.; Henry, R.; Lavigne-Cruège, V.; Dubourdieu, D. J. Agric. Food Chem. 1998, 46, 3095–3099. Tominaga, T.; Dubourdieu, D. J. Agric. Food Chem. 2006, 54, 29–33. Vermeulen, C.; Lejeune, I.; Tran, T. T. H.; Collin, S. J. Agric. Food Chem. 2006, 54, 5061–5068. Martin, N.; Neelz, V.; Spinnler, H. E. Food Qual. Prefer. 2004, 15, 247–257. Berger, C.; Khan, J. A.; Molimard, P.; Martin, N.; Spinnler, H. E. Appl. Environ. Microbiol. 1999, 65, 5510–5514. Spinnler, H. E.; Berger, C.; Lapadatescu, C.; Bonnarme, P. Int. Dairy J. 2001, 11, 245–252. Cuer, A.; Dauphin, G.; Kergomard, A.; Dumont, J. P.; Adda, J. Appl. Environ. Microbiol. 1979, 38, 332–334. Deetae, P.; Bonnarme, P.; Spinnler, H. E.; Helinck, S. Appl. Microbiol. Biotechnol. 2007, 76, 1161–1171. Helinck, S.; Le Bars, D.; Moreau, D.; Yvon, M. Appl. Environ. Microbiol. 2004, 70, 3855–3861. Lamberet, G.; Auberger, B.; Bergère, J. L. Appl. Microbiol. Biotechnol. 1997, 48, 393–397. Helinck, S.; Spinnler, H. E.; Parayre, S.; Dame-Cahagne, M.; Bonnarme, P. FEMS Microbiol. Lett. 2000, 193, 237–241. Ferchichi, M.; Hemme, D.; Nardi, M.; Pamboukdjian, N. J. Gen. Microbiol. 1985, 131, 715–723. Amarita, F.; Yvon, M.; Nardi, M.; Chambellon, E.; Delettre, J.; Bonnarme, P. Appl. Environ. Microbiol. 2004, 70, 7348–7354. Dias, B.; Weimer, B. Appl. Environ. Microbiol. 1998, 64, 3327–3331. Ganesan, B.; Seefeld, K.; Weimer, B. C. Appl. Environ. Microbiol. 2004, 70, 6385–6393. Smit, B. A.; Engels, W. J. M.; Wouters, J. T. M.; Smit, G. Appl. Microbiol. Biotechnol. 2004, 64, 396–402. 134 Qian et al.; Volatile Sulfur Compounds in Food ACS Symposium Series; American Chemical Society: Washington, DC, 2011.

Downloaded by EAST CAROLINA UNIV on November 5, 2016 | http://pubs.acs.org Publication Date (Web): August 24, 2011 | doi: 10.1021/bk-2011-1068.ch006

24. Deetae, P.; Bonnarme, P.; Spinnler, H. E.; Helinck, S. Appl. Microbiol. Biotechnol. 2007, 76, 1161–1171. 25. Tominaga, T.; Guimbertau, G.; Dubourdieu, D. J. Agric. Food Chem. 2003, 51, 1373–1376. 26. Tominaga, T.; Dubourdieu, D. J. Agric. Food Chem. 2006, 54, 29–33. 27. Sourabié, A. M.; Spinnler, H. E.; Bonnarme, P.; Saint-Eve, A.; Landaud, S. J. Agric. Food Chem. 2008, 56, 4674–4680. 28. Maciel, M. L.; Hansen, T. J.; Aldinger, S. B.; Labows, J. N. J. Agric. Food Chem. 1986, 34, 923–927. 29. Teranishi, R.; Corse, J. W.; McFadden, W. H.; Black, D. R.; Morgan, A. I. J. Food Sci. 1961, 28, 78–483. 30. Tominaga, T.; Guimbertau, G.; Dubourdieu, D. J. Agric. Food Chem. 2003, 51, 1016–1020. 31. Kolor, M. G. J. Agric. Food Chem. 1983, 31, 1125–1127. 32. Ettre, L. S.; Welter, C.; Kolb, B. Chromatographia 1993, 35, 73–84. 33. Arfi, K.; Landaud, S.; Bonnarme, P. Appl. Environ. Microbiol. 2006, 72, 2155–2162. 34. Lowe, P. N.; Hodgson, J. A.; Perham, R. N. Biochem. J. 1983, 215, 133–140. 35. Vallet, A.; Lucas, P.; Lonvaud-Funel, A.; de Revel, G. J. Appl. Microbiol. 2008, 104, 1833–1840. 36. Dickinson, J. R.; Dawes, I. W. J. Gen. Microbiol. 1992, 138, 2029–2033. 37. Malcorps, P.; Dufour, J. P. Eur. J. Biochem. 1992, 210, 1015–1022.

135 Qian et al.; Volatile Sulfur Compounds in Food ACS Symposium Series; American Chemical Society: Washington, DC, 2011.